The overarching goal of this proposal is to use evidence-based methodology to evaluate Magnetoencephalography (MEG) as a clinical tool for evaluation and planning of surgical treatment for patients with epilepsy. In patients where interictal spikes or ictal activity are not detectable, we will evaluate protocols for localizing epileptic cortex during brief MEG exams, by 1) localizing abnormal high frequency oscillations and slow waves using advanced analysis methods (LCMV beamformer and MR-FOCUSS);and 2) withdrawal of a short-acting barbiturate or 3) withdrawal of vagus nerve stimulation. We will validate these advanced MEG analysis techniques and the existing single equivalent current dipole (ECD) method against electrocorticography (ECoG), and determine their sensitivities and specificities. To benefit patients with complex patterns of partial epilepsy, we will use MEG to study the timing and propagation of spikes, and use diffusion tensor imaging (DTI) tractography to identify the underlying network to distinguish multifocal epilepsy from unifocal epilepsy with consistent patterns of propagation. We will determine if MEG, alone, is sufficient to replace invasive intraoperative mapping for intractable epilepsy by comparison of simultaneous MEG and ECoG localizations of interictal spikes. Similarity and distribution of spike waveforms observed at the locations of the ECoG electrodes will be compared to MEG signals after a spatial filter (LCMV beamformer) has been applied. We will also validate the sensitivity and specificity of functional MEG language and memory imaging as a noninvasive alternative to the intracarotid amobarbital procedure (IAP, Wada test) by demonstrating that the spatial and temporal resolution provided by MEG is necessary to image the sequence and evolution of cortical source activity involved in language and memory;specifically 1) that MEG language and memory laterality indices agree with those from the Wada test;2) that MEG language and memory imaging agrees with direct electro-cortical stimulation mapping (DECS);and 3) that MEG language imaging protocols are reliable (testing intrasubject reproducibility). We will determine the sensitivity and specificity of advanced analytical methods by comparison with the established ECD fit (when feasible) and semiology to electrocorticography. We will accomplish these goals by means of examining a set of readily testable hypotheses. Thus the completion of the proposed studies will establish MEG as a useful and reliable diagnostic tool. PUBLIC HEALTH RELEVANCE: The overall goal of this proposal is to extend MEG as a clinical tool for evaluating and planning treatment for patients with epilepsy. The objective of this study is to establish MEG as a reliable clinical diagnostic tool.